Databricks Lakebase vs Neon

Summary

• Neon and Lakebase are not competing products. Neon joined Databricks in May 2025, and Neon's own site states that the same technology behind Neon powers Lakebase.
• Lakebase is a fully managed Postgres database integrated into the Databricks platform, built for real-time transactional applications that run alongside lakehouse data.
• Lakebase Postgres is also available on the Neon.com platform, where it sits next to Auth, Object Storage and Functions for developers and agent platforms.
• Both document the same core database features: autoscaling, scale to zero, branching, instant restore and read replicas. The difference is the platform around the database: Unity Catalog and the lakehouse on Databricks, backend services on Neon.com.
• On Databricks, Lakebase connects to lakehouse data through synced tables, the Lakebase Change Data Feed (Public Preview) and registration in Unity Catalog.

Databricks Lakebase vs Neon

Neon and Lakebase aren't rivals. Neon joined Databricks in May 2025, and the same technology behind Neon powers Lakebase. The practical difference is where you run it. Lakebase is a fully managed Postgres database integrated into the Databricks platform. Neon.com is a developer backend platform, and Lakebase Postgres is its database.

What is Neon?

Neon's documentation opens with "Neon is the backend for apps and agents." Its database is Lakebase Postgres, described as serverless Postgres with branching, autoscaling, scale to zero and instant restore. Neon also lists Auth (Managed Better Auth), Object Storage (S3-compatible, branches with your database) and Functions.
The timeline is short. Databricks announced its agreement to acquire Neon on May 14, 2025. On June 11, 2025, Databricks launched Lakebase as a fully managed Postgres database powered by Neon technology.

What is Lakebase?

The Databricks docs describe Lakebase as a fully managed Postgres database integrated into the Databricks platform, for building real-time transactional applications alongside your lakehouse data. Applications connect with standard Postgres clients and tools.
Lakebase is generally available on AWS (since January 22, 2026) and on Azure (since March 2, 2026), and it is in Beta on Google Cloud. The docs also list high availability, with automatic failover to keep the database available during compute failures.

Lakebase vs Neon: how they compare

Aspect	Lakebase on Databricks	Neon (neon.com)
What it is	A fully managed Postgres database integrated into the Databricks platform	"The backend for apps and agents", with Lakebase Postgres as the database
Platform around the database	Unity Catalog governance, lakehouse analytics, notebooks and AI workflows	Backend services for developers, startups and agent platforms: Auth, Object Storage, Functions
Autoscaling	Compute adjusts to workload demand	Compute adjusts to the current load, with no manual intervention or restarts
Scale to zero	Inactive computes are suspended and reactivate on the next query	Compute is suspended after inactivity and reactivates when queried
Branching	Copy-on-write branches for development, testing and staging	A branch is a copy-on-write clone of your data
Instant restore	Create a branch from any point in your history window	Restore a root branch to an earlier point in its history
Read replicas	Read-only computes on the same storage, eventually consistent	Read replicas on the same storage, eventually consistent
Clouds	AWS and Azure (generally available), Google Cloud (Beta)	AWS regions (Neon's docs mark its Azure regions as deprecated)

How Lakebase connects to the lakehouse

Lakebase tables are Postgres tables. Delta tables enter the picture through three documented integrations:

• Synced tables. Unity Catalog tables sync into Lakebase so applications can query lakehouse data with low latency. Supported sources are managed and external Delta tables, managed and external Iceberg tables, views and materialized views. The sync modes are Snapshot, Triggered and Continuous.
• Lakebase Change Data Feed (Public Preview). Each insert, update and delete on a Postgres table is captured from the write-ahead log and stored as a row in a Unity Catalog managed Delta table.
• Registration in Unity Catalog. Registering a Lakebase database creates a read-only catalog that represents your Postgres database. You query it with a serverless SQL warehouse, and Unity Catalog permissions, lineage and audit logs apply.

Which one should you use?

Use Lakebase on Databricks when the application database should sit next to lakehouse data that Unity Catalog governs. The docs describe four patterns: serve lakehouse data in Postgres, store Postgres changes in the lakehouse, run an application backend, and power agents and ML.
Use Neon.com when you want Lakebase Postgres together with Neon's backend services: Auth, Object Storage and Functions.

FAQs

Is Neon a competitor to Lakebase?

No. Neon is part of the Databricks platform, and the same technology behind Neon powers Lakebase.

Is Lakebase built on Delta Lake?

No. Lakebase is a Postgres database with Postgres tables. Delta tables connect through synced tables (Unity Catalog to Lakebase), the Change Data Feed (Postgres to Delta, Public Preview) and registration in Unity Catalog.

Does Unity Catalog govern every connection to Lakebase?

No. Unity Catalog permissions control query access through SQL warehouses on a registered Lakebase catalog. Direct connections to the database use Postgres roles and permissions independently.

Can I query Lakebase from a SQL warehouse?

Yes, after you register the database in Unity Catalog. The registered catalog is read-only and needs a serverless SQL warehouse. Pro and Classic SQL warehouses are not supported.

Is Lakebase Postgres only available inside Databricks?

No. The Databricks docs note that Lakebase Postgres is also available on the Neon.com platform and point to the Neon documentation for using it there.